This proposal Spectral-Spatial RF Pulses for Gradient Echo MRI is an MRI technology development project to design spectral-spatial Radio Frequency (RF) excitations using single and multiple transmitters at 3T. These pulses will be designed with the goal of suppressing unwanted lipid signal, reducing susceptibility artifacts, and improving slice profile (B1+) uniformity in gradient echo MRI. Gradient echo applications such as blood oxygen level dependent (BOLD) brain functional MRI (fMRI) are plagued by large signal voids in the inferior brain regions due to magnetic susceptibility variations. Furthermore, the high fields required for good gradient echo contrast make the images prone to intensity variations from B1+ inhomogeneity. Methods that address these limitations are important to exploit the full benefits of MRI for improved health care and research. We first propose 2D spectral spatial pulses for slice and frequency selectivity on one or multiple transmitters. These pulses can be used for both lipid suppression and the cancellation of the through-plane susceptibility gradient. The susceptibility artifact correction assumes that the through-plane gradient is a function of off- resonance frequency. Acquiring field maps to determine the spatial distribution of through-plane gradients and off-resonance will test this assumption. The spatial variations of the maps will then be exploited using parallel transmission methods. The next approach will be to design 4D spectral-spatial pulses for parallel transmitters to develop excitations that simultaneously correct for through-plane susceptibility artifact, in-plane transmitter (B1+) inhomogeneity, and provide lipid suppression. The pulse generation algorithms will then be ported for use on graphics programming units (GPUs) for increased speed. The pulses will be tested and characterized with simulations and phantom and human control gradient echo imaging studies. Final validation of the pulses will use human control scanning with susceptibility weighted imaging (SWI), T2* mapping, and breath-holding BOLD fMRI experiments. Success in developing the methods described in this proposal will overcome major limitations in gradient echo MRI, making feasible a broad range of clinical applications not previously possible. Furthermore, the application spectral-spatial pulses and parallel transmitters is novel to this proposal and represent a big step forward in multi-dimensional RF pulse design.